About every booth counts

Publication status

Private preview — unannounced founder-review build. A search-excluded Cloudflare Pages preview exists for founder review. This is not a public launch. A dataset release or any outreach still requires separate, explicit founder approval and has not happened.

Method

• Roster rule: Status in {Current, Appointment}. Abolition-status rows are excluded.
• Categories: Full / Assisted / None map exactly to the AEC’s own WheelchairAccess field values. Blank or unrecognised values are disclosed as “Not stated” — never guessed or silently dropped.
• Minimum-n handling: A division is flagged low_n when its total roster count is below the threshold; divisions are never suppressed here (the smallest, Lingiari, still has 17 booths), but every category is always presented as a literal 'k of n' count alongside its percentage so a small numerator (e.g. Fairfax's 1 of 31 Full) is never shown as a bare, seemingly precise percentage on its own.
• No invented figures: this site counts polling places by the AEC’s own category. It does not estimate how many people are affected, and it does not compute an accessibility score, index, or rating of any kind.

Why there is no 2022-to-current trend shown

Status: closed.

The 2022 baseline and current snapshot are both computed and stored below, but this build deliberately does not compute or publish a 2022-to-current change, delta, or trend figure. The national Full share appears to shift from roughly 21.5% (2022) to roughly 42% (current), which is a much larger move than the AEC's own public wording of 'around 20% increase' -- suggesting a possible assessment-framework or category change between snapshots rather than a real-world change of that size. Per the compact (PLAN.md Non-negotiable #7 and never-do #9), no trend line, trend chart, or cross-year change claim ships until a methodology audit (docs/trend-audit.md) resolves this or explicitly break-marks it.

Audit doc: docs/trend-audit.md. Both years’ literal figures are computed and retained in the underlying dataset for future audit use; this site simply never presents them side-by-side as a change or trend while the gate is closed.

Framing commitments

• No accessibility “score,” index, or rating — only literal AEC-category counts and percentages.
• No ranking or division-vs-division comparison table anywhere on this site. The only comparative surface is the single national distribution shown on the index page; division pages never compare one division’s figures to another’s.
• No distance-in-kilometres headline, ever. Where a polling place is rated None, we report that as a literal count in words. If a None-rated venue is a voter’s only practical local option, we say that plainly in a sentence — we do not turn it into a raw kilometre figure, and we do not name or shame the venue.
• No venue-level callouts, worst-lists, or venue naming as a failure. Any accountability framing here is at the division level only.
• Candidate/party-neutral copy always. No party names, candidate names, endorsements, or voting suggestions anywhere on this site.

Rights and attribution

Source data: Australian Electoral Commission (AEC) per-polling-place wheelchair-access CSVs. © Commonwealth of Australia, licensed under CC BY 4.0. This project is not affiliated with the AEC. Not an official AEC product, and not endorsed by the AEC. During elections the AEC's own tool is the canonical source — this is historical/archival. — during any federal election period, use the AEC’s own official website (aec.gov.au) for current polling-place information.